Open AccessUSING JUST NOTICEABLE DIFFERENCES TO INTERPRET TEST OF SPOKEN ENGLISH SCORES
Author(s) -
Stricker Lawrence J.
Publication year - 1997
Publication title -
ets research report series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.235
H-Index - 5
ISSN - 2330-8516
DOI - 10.1002/j.2333-8504.1997.tb01725.x
Subject(s) - test (biology) , psychology , just noticeable difference , standard deviation , statistics , audiology , significant difference , concurrent validity , mathematics , internal consistency , psychometrics , computer science , artificial intelligence , medicine , paleontology , biology
ABSTRACT This study explored the value of obtaining a Just Noticeable Difference (JND) — the difference in scores needed before observers discern a difference in examinees' English proficiency – for the current Test of Spoken English (TSE®) as a means of interpreting scores in practical terms, using college students' ratings of their international teaching assistants' English proficiency and adapting classical psychophysical methods. The test's concurrent validity against these ratings was also appraised. Three estimates of the JND were obtained. They varied considerably in size, but all were substantial when compared with the standard deviation of the TSE scores, the test's standard error of measurement, and guidelines for the effect size for mean differences. The TSE test correlated moderately with the rating criterion. The JND estimates appear to be meaningful and useful in interpreting the practical significance of TSE scores, and the test has some concurrent validity.